Implicit conversion from float -135.0879249706917 to int loses precision